From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 17622 invoked by alias); 5 Apr 2005 07:51:42 -0000 Mailing-List: contact gdb-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sources.redhat.com Received: (qmail 16877 invoked from network); 5 Apr 2005 07:51:19 -0000 Received: from unknown (HELO cat.co.za) (196.33.33.51) by sourceware.org with SMTP; 5 Apr 2005 07:51:19 -0000 Received: (qmail 23546 invoked by uid 1002); 5 Apr 2005 08:08:21 -0000 Received: from stuartb@cat.co.za by mail.cat.co.za by uid 807 with qmail-scanner-1.21 (clamscan: 0.70. Clear:RC:1(196.33.33.52):. Processed in 1.723848 secs); 05 Apr 2005 08:08:21 -0000 Received: from unknown (HELO STU95) (196.33.33.52) by 0 with SMTP; 5 Apr 2005 08:08:19 -0000 Message-ID: <108401c539b4$2430c680$5e22a8c0@STU95> From: "Stuart Brooks" To: "Stuart Brooks" , "Nathan J. Williams" Cc: References: <0f4601c538fc$30c35970$5e22a8c0@STU95> <0fb501c539a1$e53857c0$5e22a8c0@STU95> Subject: Re: gdb 5.3nb1 crashing when accessing elements in static libs Date: Tue, 05 Apr 2005 07:51:00 -0000 Organization: Intervid Technologies MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-SW-Source: 2005-04/txt/msg00037.txt.bz2 > > I'm assuming this is i386? You're going to be in some pain with C++ on > > this combination; netbsd/i386 2.0 sets gcc to use the stabs debugging > > format by default instead of dwarf-2, because of a bug in 3.3's dwarf > > generation (netbsd PR #25094; equivalent to GCC bug #16788) [I'm not > > at all sure that this is the right tradeoff for that bug]. > > Yes, this is i386 :( What would you recommend I do? I suppose I could > install a later version of gcc (seems 3.4.0 has this fixed) but I would > imagine that that will need a full system rebuild depending on the > compatibility between 3.3.3 and 3.4.0. Or is there some easier solution, > maybe setting g++ to use dwarf-2 and accepting that the offset for long > long structs will be incorrect (is this a train smash?)? Compiling everything with -gdwarf-2 gives me the ability to access symbols in classes compiled in static libraries which is great (I couldn't do this before). An interesting thing though is that all global variables seem to be viewed/treated as ints. eg. // global variables MyClass* g_myclass_p; // initialized later MyClass g_myclass; int* g_int_p=(int*)2000; int g_int=1234; will give (in gdb printouts): g_myclass_p = 13423223 g_myclass = 0 g_int_p = 2000 g_int = 1234 Again, this is a new one for me. Does anyone know if this is somehow related? Thanks for the help so far, I really appreciate it! - Stuart